We re-examine the deuteron elastic breakup cross sections on 12 C and 10 Be at low incident energies, for which a serious discrepancy between the continuum-discretized coupled-channels method (CDCC) and the Faddeev-Alt-Grassberger-Sandhas theory (FAGS) was pointed out. We show the closed-channels neglected in the preceding study affect significantly the breakup cross section calculated with CDCC, resulting in good agreement with the result of FAGS.PACS numbers: 24.10. Eq, 25.60.Gc, 27.20.+n Introduction. Projectile breakup reactions have played a major role in studying the structure of loosely-bound nuclei [1]. Such a reaction contains at least three particles in the final state. Thus, one may say that the accurate description of the three-body breakup process is a minimum requirement for nuclear reaction theories. It is well known that the Faddeev theory [2], or, alternatively, the Alt-GrassbergerSandhas (AGS) theory [3] gives the exact solution to such a three-body scattering problem. On the other hand, the continuum-discretized coupled-channels method (CDCC) [4][5][6] has widely been applied with high success to projectile breakup reactions at various incident energies. The theoretical foundation of CDCC was given in Refs. [7,8] in connection with the distorted-wave Faddeev formalism [9]. Quite recently [10], invention of the treatment of the Coulomb interaction made the Faddeev-AGS theory (FAGS) applicable to various three-body breakup reactions, and the results of FAGS have directly been compared with those of CDCC. In many cases the two give very similar cross sections, which validates CDCC as an effective three-body reaction model, as predicted in Refs. [7,8].